The Characteristics and Regulatory Mechanisms of Superoxide Generation from eNOS Reductase Domain

In addition to superoxide (O2 .-) generation from nitric oxide synthase (NOS) oxygenase domain, a new O2 .- generation site has been identified in the reductase domain of inducible NOS (iNOS) and neuronal NOS (nNOS). Cysteine S-glutathionylation in eNOS reductase domain also induces O2 .- generation from eNOS reductase domain. However, the characteristics and regulatory mechanism of the O2 .- generation from NOS reductase domain remain unclear. We cloned and purified the wild type bovine eNOS (WT eNOS), a mutant of Serine 1179 replaced with aspartic acid eNOS (S1179D eNOS), which mimics the negative charge caused by phosphorylationand truncated eNOS reductase domain (eNOS RD). Both WT eNOS and S1179D eNOS generated significant amount of O2 .- in the absence of BH4 and L-arginine. The capacity of O2 .- generation from S1179D eNOS was significantly higher than that of WT eNOS (1.74:1). O2 .- generation from both WT eNOS and S1179D eNOS were not completely inhibited by 100nM tetrahydrobiopterin(BH4). This BH4 un-inhibited O2 .- generation from eNOS was blocked by 10mM flavoprotein inhibitor, diphenyleneiodonium (DPI). Purified eNOS reductase domain protein confirmed that this BH4 un-inhibited O2 .- generation originates at the FMN or FAD/NADPH binding site of eNOS reductase domain. DEPMPO-OOH adduct EPR signals and NADPH consumptions analyses showed that O2 .- generation from eNOS reductase domain was regulated by Serine 1179 phosphorylation and DPI, but not by L-arginine, BH4 or calmodulin (CaM). In addition to the heme center of eNOS oxygenase domain, we confirmed another O2 .- generation site in the eNOS reductase domain and characterized its regulatory properties.


Introduction
Endothelial nitric oxide synthase (eNOS, type III) is an important enzyme that is involved in a variety of fundamental functions such as cardiovascular tension regulation, proliferation of vascular smooth muscle cells, leukocyte adhesion and platelet aggregation [1][2]. The eNOS gene undergoes co-translational N-myristoylation and post-translational cysteine palmitoylation in the nuclear unit and Golgi, respectively [3][4]. Endothelial NOS (eNOS, type III) contains two functionally independent domains: an N-terminal oxygenase domain and a Cterminal reductase domain. In the eNOS oxygenase domain, there are sites that bind BH4 and L-arginine and convert L-arginine into L-citrulline, generating nitric oxide (NO). In the eNOS reductase domain, there are FMN and FAD/NADPH binding sequence and they are structurally and functionally similar with NADPH-cytochrome P-450 reductase. Between the oxygenase and reductase domains, there is an amino acid sequence binds to calmodulin (CaM). This region is necessary for maintaining the eNOS dimer structure and normal function.
Early in vitro studies demonstrated that all three kinds of NOS produced superoxide in Larginine and BH 4 -depleted condition [5][6][7]. Functional studies showed that superoxide (O 2 .-) formation from eNOS occurred primarily at the heme center of its oxygenase domain. This O 2 .synthesis from eNOS requires Ca 2+ /CaM and is primarily regulated by BH 4 rather than Larginine [8]. The eNOS reductase domain contains 684 amino acids and there are two major subdomains: a FAD/NADPH binding sequence (758-1004), which generates the electrons and delivers them to the heme center of oxygenase domain of another monomer; two FMN binding sequences (from 522 to 705), which connect with the CaM binding sequence [9]. Glutamine 894, a reductase domain amino acid, polymorphisms changed eNOS response to L-argnine and further affected eNOS activity [10]. Recent study showed that S-glutathionylation of cysteines in eNOS reductase domain reversibly decreased eNOS activity and increased O 2 .generation. However, whether these cysteines glutathionylation-mediated O 2 .generation has not yet been fully investigated [11]. Additionally, a iNOS reductase domain subcloning study demonstrated that the FAD/NADPH and FMN binding sequence was another superoxide generation site in iNOS [12]. As all three isoforms of NOSs have FMN and FAD/NADPH binding sequence in reductase domains, it would be interesting to determine the O 2 .generation location and regulatory properties in eNOS reductase domain. ENOS activity is regulated by cofactors as well as several critical amino acid modifications [13]. For example, the phosphorylation target of Akt is bovine Serine 1179 (Serine 1177 of human eNOS). Serine 1179 phosphorylation results in enhancement of eNOS activity and increased sensitivity to Ca 2+ and CaM. Mutation of S1179 to aspartic acid (S1179D), which mimics the negative charge caused by phosphorylation, also increases eNOS activity [14][15]. Additionally, cytochrome C reduction studies demonstrate enhanced electron flux through the reductase domain of S1179D bovine eNOS compared to wild type eNOS [16]. Early in vitro studies demonstrated that heat shock protein 90 (HSP90), and not Serine 1179 phosphorylation, modulates O 2 .generation in endothelial cells using DMPO spin trapping [17].
However, recent studies indicate that O 2 .production in EC is regulated by eNOS Serine 1179 phosphorylation through two independent pathways: a direct pathway and an indirect pathways [18][19][20][21].
In this study, we utilized a stable spin trap probe (DEPMPO) with EPR to investigate whether the eNOS reductase domain has the potential to generate superoxide. We also aimed to determine if the classic eNOS cofactors (BH4, L-arginine and Ca 2+

ENOS Purification
Both bovine WT eNOS and S1179D eNOS pcDNA 3 plasmid were used as templates and performed PCR. The forward primer was: 5'-CGGAATTCAACATGGGCAACTTGAAGAGTGTG GGCCAG-3' and the backward primer is 5'-GCTCTAGATCATCAGGGGCCGGGGGTGTCTGG-3'. The bovine eNOS reductase domain (521-1205) PCR forward primer was: 5'-CGGAATTC AACAAAGCAACCATCCTGTACG-3' and the backward primer is: 5'-GCTCTAGATCA TCAG GGGCCGGGGGTGTCTGG-3' (start codon and stop codon was underlined). The PCR products were digested with EcoRI and XbaI and the product was subcloned into pCW plasmid [7,[22][23]. The resulting pCW plasmid was transformed into E.Coli (BL21) and cultured overnight. After harvesting the E.Coli by centrifugation, the bacteria pellet was homogenized in buffer A, which contains 50 mM TrisÁHCl, pH 7.6, 0.1 mM EDTA, 150mM NaCl, 0.1mM DTT, 10% glycerol and protease inhibitor cocktail solution (Sigma). After being centrifuged (16,000 x g, 10 min) at 4°C, the supernatant was applied to a 2',5'-ADP-Sepharose 4B column pre-equilibrated in buffer A. The column was extensively washed with buffer A containing 600 mM NaCl and TrisÁHCl buffer (50 mM, pH 7.6). The protein was then eluted with 5 mM AMP in 50 mM TrisÁHCl (pH 7.6). The eluate was washed and concentrated using Centricon-100 (Amicon) concentrators. Further purification was performed using FPLC. The final purified WT and S1179D eNOS proteins were isolated from the fractions which had high absorption at 280nm and 408nm. Protein content of the preparations was assayed with the Bradford reagent (Bio-Rad) using bovine serum albumin as standard. The purity of eNOS was determined by SDS-PAGE and visualized with Coomassie brilliant blue (R-250, Bio-Rad) staining. The purified eNOS preparations exhibited one prominent band on gels with a molecular mass of 135 kDa. Purified eNOS samples were stored in 50 mM TrisÁHCl (pH 7.6) buffer with 10% glycerol at -80°C.

EPR Spectroscopy and Spin Trapping
Spin-trapping measurements of oxygen free radicals were performed in 50 mM Tris-HCl buffer, pH 7.6, containing 0.5 mM NADPH, 0.5 mM Ca 2+ , 10 μg/ml calmodulin, 400nM purified eNOS, and 20 mM spin trap DEPMPO. EPR spectra were recorded in a disposable micropipette (50μl, VWR Scientific) at room temperature (23°C) with a Bruker EMX spectrometer operating at X-band with a high sensitive (HS) cavity (Bruker Instrument, Billerica, MA)using a modulation frequency of 100 kHz, modulation amplitude of 0.5 G, microwave power of 20mW, and microwave frequency of 9.863GHz as described [12,25]. The central magnetic field was 3510.0 Gauss (G) and the sweep width was 140.0 G. The time constant was 163.84 ms, the sweep rate was 40.96 ms, and the receiver gain was 2×10 6 .

NADPH Consumption by eNOS
NADPH oxidation was measured spectrophotometrically at 340 nm [26]. The reaction systems were the same as described in EPR measurements, and the experiments were run at room temperature. The rate of NADPH oxidation was calculated using a molar extinction coefficient of 6.22 /mM/cm Statistics Data are expressed as means ± SE. Comparisons were made using a two-tailed paired or unpaired Student's t-test. Differences were considered to be statistically significant at P < 0.05.

Expression of eNOS and Activity Analysis
Both wild type and S1179D eNOS were expressed and purified from E. Coli. In a culture of 2 liters, approximately 3.0-4.0 mg of eNOS was typically recovered using 2'5'-ADP Sepharose 4B chromatography. Further purification was performed using FPLC. The final purified products were identified by SDS-PAGE gel ( Fig 1A). Previous studies demonstrated that eNOS has higher activity after serine 1179 phosphorylation or Serine 1179 mutation into aspartic acid (S1179D eNOS) [14,27] affect the DEPMPO-OOH adduct signal generated from XO/X system (Fig 2A left middletrace), which was consistent with previous observation [30]. Compared to XO/X system, 4μg eNOS generated similar density of DEPMPO-OOH adduct signal in the absence of L-arginine and BH4 (Fig 2A right top trace and Fig 2B). In contrast to DEPMPO-OOH adduct signaling generated from XO/X system, DEPMPO-OOH adduct signaling from eNOS was significantly inhibited by 100nM BH4 (Fig 2A right middle trace and Fig 2B). BH4 did not non-specifically scavenge the O 2 .until the concentration reached 500nM (data not shown). Moreover, DEPM-PO-OOH adduct signals from both the XO/X system and from eNOS were significantly   (Fig 2A bottom traces and Fig 2B). NADPH center, which locates in the reductase domain of eNOS [11]. As WT eNOS and S1179D eNOS were treated with 10mM DPI, a flavoprotein inhibitor, and the DEPMPO-OOH signals significantly decreased (Fig 4A bottom traces and 4B).

L-Arginine, BH 4 Did Not Affect the O 2 .-Generation from eNOS Reductase Domain
To exclude the potential effect of the oxygenase domain on the O 2 .release from reductase domain, we used 1mM L-NAME, a heme center inhibitor. L-NAME did not cause a significant inhibitory effect on O 2 .generation from WT eNOS or S1179D eNOS when compared to control. The DEPMPO-OOH adduct signals from WT eNOS and S1179D eNOS were 37661± 575.2 and 53783 ± 2958.5, respectively. When WT eNOS and S1179D eNOS were treated with L-arginine (1mM), a substrate for NO generation of eNOS which also binds in heme center, system and eNOS were completely inhibited by SOD (200 units/mL). The EPR spectra were representative spectra from five independent experiments and the density was presented as Mean ±SD. doi:10.1371/journal.pone.0140365.g002 the DEPMPO-OOH adduct signals generated were 38367±4841.58 and 47520±1533.9 respectively, with no significant difference when compared to controls (Fig 5A). Early studies showed that BH4 and L-arginine were the determinant factors regulating NO or O 2 .generation from eNOS heme center [32][33]. Next we investigated whether BH4 or BH4 and L-arginine together affected O 2 .release from the reductase domain. WT eNOS and S1179D eNOS were incubated BH4 (100nM) at a Ca 2+ /CaM depletion condition. As shown in Fig 5B, the DEPMPO-OOH signals generated from WT eNOS and S1179D were 47930.4 ±4679 and 56564.81±5964.4, respectively. When WT eNOS and S1179D eNOS were incubated with BH4/ L-arginine (100nM/1mM), the DEPMPO-OOH signals generated from WT eNOS and S1179D were 41459.65±3403.8 and 54607.8±50.7, respectively ( Fig 5B). There was no significant difference between BH4 treated groups and BH4/ L-arginine groups in the absence of Ca 2+ and CaM.

The Affirmative Evidence of O 2 .-Generation from the eNOS Reductase Domain
To further determine the mechanism of O 2 .release from the eNOS reductase domain, we subcloned the bovine eNOS reductase domain from WT eNOS plasmid and expressed it in E.Coli. 1μg purified eNOS reductase domain protein was identified on the SDS-page gel (Fig 6A).     Fig 7).

Discussion
In this study, we obtained three important findings. . Although eNOS, nNOS and iNOS have different oxygenase domains and enzyme activities, they have similar structure in their reductase domains those function like NADPH-cytochrome P-450 [13]. Previous studies showed that electron flux leaked from iNOS reductase domain and was not affected by L-arginine [12]. Also, nNOS HSP90 binding studies showed that HSP90 could largely but not completely inhibit O 2 .generation from nNOS [34][35] (Fig 4). NADPH consumption rates also confirm that eNOS Ser1179 phosphorylation enhances electron generation and electron delivery efficiency, which is indicative of more oxygen molecules converting into O 2 .-. We are on the way to determining the importance of electron redistribution between the eNOS oxygenase domain and eNOS reductase domain caused by eNOS Ser1179 phosphorylation. Previous observations have found that the eNOS reductase domain binds with NADPH and generates an electron, which is delivered into the heme center of the oxygenase domain [32].